Hi Toshi Hi Cris folks Thanks for your interst in the matter ... Am following the siutuation
> Srinath, if you add a MU header on A1#handleRequest() with a > role as 'next', the node couldn't be distinguished whether it is > sure for A2 or not by A2#HandleRequest(). > I think that the 'next' of role-attribute has a special meaning > what the attribute is for the *NEXT* SOAP node (i.e. not for the > current SOAP node, = not for other handlers behind A1). > But, A2#HandleRequest() has no way to know the 'next' is added > by A1#handleRequest() or the right predecessor as an intermediary. > Thus, A1#HandleRequest() shouldn't add a MU header as a role 'next'. > If you strictly want to pass the node to the next intermediary, > you should add the node on A1#handleResponse(). thanks Cris I belive this is the case. Thanks for your time regards Srianth On Wed, 20 Aug 2003 21:24:50 -0700 (PDT), Toshiyuki Kimura wrote > Hi Chris, Srianth, and folks, > > It seems to me that we reache a consensus on a part of this > matter. Chris, and I have made an equivalent response to Srianth, > despite with two different point of views as "Acutial condition > with Axis implementation" and "Idealized vision to accord with > several specifications (i.e SOAP 1.1, 1.2, and JAX-RPC)". > > In JAX-RPC....... > > If you have two JAX-RPC handler, below is the workflow. > > +---+ > | A |-->A1#handleRequest()----->A2#HandleRequest()------+ > | x | Endpint > | i | V > | s |<--A1#handleResponse()<----A2#HandleResponse()<----+ > +---+ > > You may add a header on A1#handleRequest() or A1#handleResponse(). > > Srinath, if you add a MU header on A1#handleRequest() with a > role as 'next', the node couldn't be distinguished whether it is > sure for A2 or not by A2#HandleRequest(). > I think that the 'next' of role-attribute has a special meaning > what the attribute is for the *NEXT* SOAP node (i.e. not for the > current SOAP node, = not for other handlers behind A1). > But, A2#HandleRequest() has no way to know the 'next' is added > by A1#handleRequest() or the right predecessor as an intermediary. > Thus, A1#HandleRequest() shouldn't add a MU header as a role 'next'. > If you strictly want to pass the node to the next intermediary, > you should add the node on A1#handleResponse(). > > Could you make a sense ? > > In addition, I believe Axis is a major JAX-RPC implementation. > But, it doesn't mean that Axis perfectly supports all of spec such > as SOAP 1.1, 1.2, and JAX-RPC 1.0. The great majority of SOAP 1.1 > and JAX-RPC 1.0, and also a small part of SOAP 1.2 have coverd in > Axis of version 1.1 final. Axis will continuously work to accord > with whole of SOAP 1.2, and any others (WS-I basic profile ?). > > -- > Toshi <[EMAIL PROTECTED]> > > -----Original Message----- > From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] > Sent: Thursday, August 21, 2003 12:02 PM > To: [EMAIL PROTECTED] > Subject: Re: MustUnderstand faults > > Hi Cris > > Thanks for the posting... > What you mean is if the handler want he can make the header added > to be processed in down stream. Am quite agree with you. > > But what does SOAP spec say ... Should that handler process in the > downstream or next handler.. I feel the spec does not provide a > clear cut senario and it is to handler to decide...what to do .. > > Thanks for your time ... These part with headers seem to be so > unclear ... > the flow thought of all the people where very helpful... > > regards > Srinath > > On Wed, 20 Aug 2003 06:39:00 -0700 (PDT), Chris Haddad wrote > > In Axis....... > > > > if you place the header in the request message and don't qualify the > > actor/role, then a handler will assume that it should be processed. > > I do not believe there is a facility for a handler to distinguish > > between the original headers sent by the client and headers added > > during processing. The benefit is that a handler can transform the > > headers into internal representations that are processed downstream. > > Also, headers can be attached to augment the message context. > > > > if you place the header in the response message, then it will be > > passed to the next soap intermediary and it is clear that it should > > not be processed by at least the request flow handlers. > > > > /Chris > > > > On Wed, 20 Aug 2003 [EMAIL PROTECTED] wrote: > > > > > Hi Toshi > > > > > > what I mean as NEXT is > "http://www.w3.org/2003/05/soap-envelope/role/next"; the > > > role given in SOAP Spec. > > > > > > what I mean is the Header addes by A1 has it's role as > > > "http://www.w3.org/2003/05/soap-envelope/role/next"; then who is going > to play > > > the next role. > > > > > > Is it next HANDLER or next NODE... > > > > > > Thanks for your time ..(enthusiasum on the subject ..) :) > > > sorry for not making clear what next is .. > > > > > > regards > > > > > > Srianth -- Lanka Software Foundation (http://www.opensource.lk) Promoting Open-Source Development in Sri Lanka
